The role of the metabolite cargo of extracellular vesicles in tumor progression

Based on their biogenesis pathways, EVs are divided into two major classes—ectosomes (or microvesicles) and exosomes [5, 8, 9] (Fig. 1). The membrane budding step, a common feature in this pathway, is similar in both classes. In addition, both EV types bud away from the cytoplasm resulting in the same membrane orientation, which is identical to the orientation of the plasma membrane [12]. In the case of ectosomes, this budding step occurs outward at the plasma membrane and results in a direct release of EVs ranging from ~ 50 nm to 1 μm in diameter. In contrast, exosomes are formed as intraluminal vesicles (ILVs) through inward budding of endosomes, which develop into multivesicular bodies (MVBs). These MVBs may fuse with lysosomes for degradation or fuse with the plasma membrane resulting in the extracellular release of ILVs as exosomes (40–160 nm in diameter) [8]. The precise molecular mechanisms of EV biogenesis have only recently started to be understood. The main driver of exosomal biogenesis is the endosomal sorting complex required for transport (ESCRT), but the existence of ESCRT-independent routes has also been proven [4, 13]. Despite their different biogenesis routes, intracellular mechanisms and sorting machineries of ectosomes and exosomes partially overlap. Shared features of different EVs make it difficult to distinguish between vesicle subpopulations [14].

EVs convey numerous proteins (e.g., tetraspanins, chaperones, biogenesis factors, signaling molecules), nucleic acids (e.g., miRNA and other non-coding RNAs, mRNA, DNA), small metabolites (e.g., sugars, amino acids, vitamins), and lipids (e.g., phosphatidylserine, cholesterol, ceramide), which are all selectively packed into vesicles in a cell type-dependent manner [14]. As EVs are distinct entities of the complex intercellular communication, their molecular fingerprint depends on the quality and state of the donor cell, and it is often influenced by microenvironmental stimuli [5, 15, 16].

The mechanisms EVs used to interact with the cell surface and transfer their cargo into the target (recipient) cells are not fully understood. Literature data suggest that these mechanisms depend on the origin and type of EV as well as the target cell [14, 17].

EVs may induce a phenotypic response in the recipient cell without internalization; receptor–ligand interactions may be sufficient to elicit signal transduction. Alternatively, EVs may transfer their cargo by direct fusion with the plasma membrane, and they may also be internalized via an active endocytic process, i.e., clathrin-, caveolin-, and lipid raft-mediated endocytosis, macropinocytosis, or phagocytosis [18]. Once in the cell, intraluminal EVs may fuse with the endosomal limiting membrane to release their content into the cytoplasm and elicit phenotypic responses in the recipient cell [14, 18, 19].

Since the origin and the physical and functional characteristics of EVs are diverse, several terms have been used for EVs in the literature. The prefixes micro- and nano- refer to their size (microvesicles, microparticles, nanovesicles, nanoparticles); ecto- and exo- refer to their presence outside the cells (ectosomes, exosomes, exovesicles). Other terms, such as oncosomes and tolerosomes, indicate their origin or function, respectively [13]. Although the nomenclature is continuously evolving, the International Society for Extracellular Vesicles (ISEV) recommends the use of “extracellular vesicle” as the “generic term for particles naturally released from the cell that is delimited by a lipid bilayer and cannot replicate.” They also suggest the use of operational terms for EV subtypes that refer to their (i) physical characteristics (small or medium/large EVs), (ii) biochemical composition (CD81⁺ EVs), or (iii) conditions of release (hypoxic EVs) [1, 20]. Here, we use the terms found in the referenced articles.

Exosomes and other classes of EVs are important mediators of cell–cell communication and play an essential role in cancer biology. It has long been well known that cancer cells secrete higher amounts of EVs than healthy cells. EVs in higher numbers have been detected in the plasma of cancer patients as well as in tumor cell cultures [21, 22].

Exosomes contribute substantially to tumor progression, invasion, and metastasis by horizontally transmitting a variety of surface and signaling molecules, oncogenic proteins, and nucleic acids to target cells, thereby altering their behavior [23, 24]. For instance, locally, in the tumor microenvironment (TME), tumor-derived EVs may convey resistance to neighboring tumor cells. These EVs can also re-educate fibroblasts and mesenchymal stem cells or activate endothelial cells, thereby inducing angiogenesis. Systemically, EVs have a crucial role in immune modulation and pre-metastatic niche formation [7, 25‐28] (Fig. 2). However, communication is not unidirectional in tumors. On the contrary, a complex, systemic communication network develops in parallel with the tumor evolution [29].

As mentioned above, the characteristic molecular fingerprint of small EVs (sEV), i.e., exosomes, is not independent of the parent/donor cell status. The metabolic status of cancer cells influences exosome secretion and content. Hypoxia, starvation, and acidosis are among the typical metabolic conditions that cancer cells undergo in the TME. Notably, all of these conditions have been shown to influence not only the rate of exosome secretion, but also the molecular composition of exosomes [30].

Before launching any investigation, one must consider the complexities of working with EVs. For instance, preparations obtained using isolation procedures that target exosomes may contain other EVs as contaminants due to the overlapping physical and biomolecular features of exosomes and microvesicles. Research guidelines for EVs are provided in the Minimal Information for Studies of Extracellular Vesicles (MISEV) to support the transparency and reproducibility of EV studies [1].

A standardized method for the isolation of EVs from cell culture supernatants or body fluids (blood, urine, saliva, etc.) has not yet been established. There are several alternative approaches to isolate and purify EVs. Important factors to consider when choosing a method include the type and volume of the EV source, the target EV subtype, the target EV yield and purity of isolates, and the downstream metabolite analysis technique. Depending on the isolation method and the EV source, abundant serum proteins (albumin, globulins) and various lipoproteins (chylomicrons, HDL, LDL, and VLDL) as well as nucleic acids on the surface of EVs may contaminate the EV isolates due to their similar physical properties. These co-isolated contaminants may interfere with the particle number and size distribution measurements and mislead biomarker analyses or functional assays [10].

Previous research has shown that EV subgroups have a unique biochemistry and function [31, 32]. This is consistent with the results of Luo et al. who found that different types of vesicles from pleural effusions showed unique metabolic enrichments [33]. Isolation methods themselves may also modify EV composition. In prostate cancer cell line models, the metabolic signature varies according to the conditions of cell culture [34]. Due to the broad range of physical and chemical characteristics of distinct metabolites, it is hardly possible to quantify all metabolites using a single approach. Therefore, it is necessary to select an appropriate method for metabolite analysis. Gas chromatography-mass spectrometry (GC–MS), liquid chromatography-mass spectrometry (LC–MS), and capillary electrophoresis-mass spectrometry (CE-MS) are the most common mass spectrometric methods used for metabolomic analyses. Additionally, in recent years, improved analytical techniques have emerged, such as ion chromatography-mass spectrometry (IC-MS) to analyze highly hydrophilic compounds [35, 36] or the capillary IC-MS as a selective and specific method to analyze anionic metabolites [37], e.g., nucleotides, sugar phosphates, and organic acids. Williams and colleagues have collected and highlighted several practical pitfalls in the field of EV metabolomics research [38].

Although lipids are dominant components of EVs, the vast majority of the EV cargo studies have investigated the protein and nucleic acid content of EVs; only a few researches have analyzed the lipid or small metabolite composition [38]. In line with this observation, EV databases, such as Vesiclepedia, ExoCarta, EVpedia, or miREV, mainly contain protein, mRNA, and miRNA entries with less lipid and metabolite data [39‐42].

The composition of EVs is comparable to that of the source donor cells, but they are also enriched in certain lipids such as cholesterol, phosphatidylserine (PS), phosphatidylcholine (PC), and phosphatidylinositol (PI), suggesting that EV may operate as cell-to-cell lipid mediators [43]. From a practical and clinical standpoint, studying the metabolomics of EVs isolated from human biofluids is the most suited approach, since the metabolome of these vesicles contains a goldmine of disease biomarkers. The EV metabolome’s potential as a source for biomarkers was first demonstrated by comparative metabolomics of plasma-derived EVs from endometrial cancer patients and healthy controls, which revealed valuable differences in these two groups [44].

Lipids are a main class of biological compounds with a wide variety of structural and signaling roles. Apart from sterols, most lipids have hydrophobic side chains and polar head groups, and the rich lipid diversity is the result of different combinations of side chains and head groups. Despite their comparable molecular complexity, there is a better understanding of the function of proteins than that of lipids, which are called as the “Cinderellas” of cell biology by Muro et al. [65].

Exosomes predominantly contain lipids, including diglycerides, sphingolipids, phospholipids, and phosphoglycerolipids, and they are enriched in specific lipids, such as cholesterol, PS, PC, and PI, which may function as cell-to-cell lipid mediators [43]. Exosomes may also carry specific bioactive lipids, including leukotrienes and prostaglandins [66].

There is accumulating evidence that the lipid content of EVs and parental cells differ; for instance, elevated levels of diacylglycerols (DAG), ceramides (Cer), sphingomyelin (SM), PC, phosphatidylethanolamines (PE) and FA were detected in EVs [31, 67, 68]. In line with this, Luo et al. identified phosphoglycerolipids, sphingolipids, and glycerolipids as major differential lipid species in lEVs and sEVs when comparing TPE with MPE enrichment of specific lipid metabolites in EVs may affect the cellular function of target cells and reflect the metabolic state of parent cells [33].

Lipids have a key role in the production and biological functions of EVs [33]. Sphingolipids, such as Cer, are critical not only in the formation and release of EVs [69], but in the regulation of cell survival and inflammation as well [70]. SM, PS, PC, PI, and cholesterol may occur in four times higher amounts in EVs than in parental cells, which contributes to the increased membrane rigidity of exosomes, and their role in the recognition and internalization of exosomes [43].

Extracellular adenosine can be produced by cells, or it can be generated from extracellular ATP. Adenosine has an extremely short half-life (10 s) in the extracellular environment due to its quick uptake by cells and irreversible conversion to inosine [120, 121]. Considerable research has been conducted in the last few years on the diverse roles and associated mechanisms of extracellular adenosine signaling. Extracellular adenosine has a wide variety of effects on cell cycle control, immunoregulation, and cytokine regulation via both direct and indirect processes, eventually contributing to the development of malignant diseases [122]. Adenosine has been detected in urinary EVs from prostate cancer patients [59]. Sayner et al. have demonstrated that EVs encapsulate cAMP to offer a second messenger compartment [123]. Ludwig et al. have recently reported that exosomes from HNC squamous cell carcinoma (HNSCC) cell lines contain cAMP and adenosine as well as adenosine metabolites, i.e. inosine, hypoxanthine and xanthine. This exosomal repository of adenosine and inosine provides a unique and key pathway for the distal transport of these purines. By shielding them from the quick uptake and metabolism, exosomes can shuttle adenosine and inosine across cells, tissues, and organ systems. Ludwig and colleagues demonstrated that exosomes from HNSCC culture supernatant contain a variety of purine metabolites, the most abundant of which are adenosine and inosine. Purine metabolites, including adenosine, were much more abundant in exosomes isolated from the plasma of HNSCC patients than in exosomes obtained from normal donors. Exosomes from patients with early-stage illness and no lymph node metastases had considerably higher levels of adenosine and 5'-GMP. At the same time, exosomal levels of purine metabolites were reduced in patients with advanced cancer and nodal involvement. Decreased purine concentration in circulating exosomes may be the result of purine metabolites being primarily used for cellular maintenance and proliferation in metastatic tumor cells instead of being packaged into exosomes and exported outside of the cell. This suggests that the molecular composition of tumor-derived exosomes and circulating exosomes is quantitatively and, possibly, qualitatively different in advanced stages compared to early malignancies [124].

Clayton and colleagues demonstrated that exosomes produced by several cancer cell types have a high capacity for ATP and 5’-AMP phosphohydrolysis, which is partly attributed to the exosomal expression of CD39 and CD73 [125]. Exosomes can carry out both hydrolytic steps sequentially to convert extracellular ATP to adenosine. Exosome-produced adenosine can induce a cAMP response in adenosine A_2A receptor-positive but not A_2A receptor-negative cells.

A recently discovered pathway, the adenosine A_2B receptor-mediated signaling for exosome-induced angiogenesis contributes to the reprogramming of endothelial cells (ECs) to an angiogenic phenotype by direct interaction, and also to the reprogramming of other cell types found in the TME, such as macrophages [126]. As previously described, A_2B receptor stimulates the growth of ECs, induces angiogenesis, leads to VEGF production and upregulation of eNOS in ECs, and stimulates macrophages to release pro-angiogenic factors [127].

Tadokoro et al. showed another mechanism leading to an increase in extracellular adenosine levels, in which perforin secreted by CD8 + cytotoxic T cells disrupts the membrane of breast adenocarcinoma-derived EVs, and adenosine passively diffuses out. Adenosine from EVs acts as an immunosuppressive metabolite by binding to the adenosine receptor and inhibits perforin secretion by cytotoxic T lymphocytes [128].

Hypoxanthine, a purine derivative, is a potential intermediate in the metabolism of adenosine and also in the synthesis of nucleic acids. Glyceraldehyde 3-phosphate (G-3-P) is an intermediate in glycolysis. Luo and colleagues detected elevated levels of hypoxanthine and G-3-P in the lEVs in the MPE in comparison to lEVs of TPE [33], which may indicate accelerated glycolysis and nucleic acid formation.

However, Ronquist et al. reported that human seminal prostasomes contain glycolytic enzymes [129]. They detected the full set of glycolytic enzymes in PCa cell-derived exosomes and observed that both types of vesicles were capable of producing ATP when substrates were available [130]. Moreover, they reported a marked distinction between the high ATPase activity of prostasomes and the low ATPase activity of a malignant cell (PC3)-derived exosomes, which leads to a larger net gain of ATP in these latter exosomes. In contrast to prostasomes, the net ATP gain of metastatic PCa cell line exosomes was considerable due to their downregulated ATPase activity. This group also found that normal and prostate cancer cells uptake EVs (prostasomes and PC3 exosomes) in an energy-dependent manner. This uptake mechanism required a continuous glycolytic flux and extracellular ATP production by EVs and/or intracellularly by recipient cells in conjunction with the presence of a functioning vacuolar-type H( +)-ATPase (V-ATPase) [130].

Few additional compounds in HNC serum-derived EVs were found to be markedly downregulated compared to healthy controls. These include citric acid, 4-hydroxybenzoic acid, and propylene glycol, while 1-hexadecanol was markedly upregulated. A few other metabolites, 1,1-dimethoxyheptane, oxoadipic acid, paramethadione could only be detected in EVs, but not in whole serum samples [45].

Folate (B9 vitamin) as a cancer-associated metabolite was identified in greater amounts in CTCL and in PCa cell line-derived EVs compared to control EVs [48]. Along with folate, overexpression of pantothenic acid (B5), niacin (B3), thiamine (B1), and pyridoxine (B6) may boost one-carbon metabolism directly or indirectly through their roles as coenzymes, hence promoting cancer development [131]. One-carbon units are required for nucleotide synthesis, methylation, and reductive metabolism, all of which contribute to the rapid proliferation of cancer cells. Also in cell line-derived EVs from PCa and CTLC, Palviainen et al. have detected elevated levels of creatinine [48].

In order to explore the relationship between metabolites found in the literature and the pathways potentially involved, we performed pathway analyses (Fig. 4). This involved a total of 62 metabolites obtained from the processed literature, and 28 of these metabolites were strongly associated with 15 different pathways. These pathways are mainly those related to AAs, but significant associations were also found for glutathione, glyoxylate, dicarboxylate metabolism, TCA cycle, pantothenate and CoA biosynthesis.

In many cases, metabolites in the same column are subsequent steps in a metabolic pathway based on MetaboAnalyst. For example, in arginine biosynthesis glutamate is converted into ornithine in four steps, which in turn is converted into arginine in two steps (with the addition of aspartate). During the reaction, fumarate is released as a side product. The metabolism of glutathione serves as another example, in which glutathione is synthesized using cysteine, glutamate and glycine in subsequent steps, and then reacts with ornithine and spermidine to form trypanothione. With regard to carbohydrate metabolism, succinate, fumarate and citrate are associated with the TCA cycle, in which these molecules act as both substrates and products.

These results show a predominance of pathways associated with AA metabolism, but this analysis has serious limitations. For example, the metabolites from different sources were examined using different methods. The outcome of the pathway analysis is strongly influenced by the fact that most functional metabolomics studies have explicitly focused on AAs, resulting in a high number of AAs and their associated pathways. Nevertheless, the results suggest that non-AA metabolites transported by EVs may also play a role in amino metabolism. In addition, the results reveal that certain metabolites, such as glutamine/glutamate, may exert a wide range of effects on the metabolism of recipient cells entering a number of pathways.